The purpose of this project is to investigate the nature of semantic category representation, and its effects on category-based reasoning. A series of experiments were conducted to test a model which makes a distinction between intrinsic and extrinsic features. The model successfully predicted performance in tasks assessing: 1) membership gradient, 2) class inclusion, and 3) linguistic contrast.